linux-kernel - [PATCH] PCI: iproc: restrict multi-MSI to single core CPUs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20210605171736.15755-1-sbodomerle@gmail.com>
Date:   Sat,  5 Jun 2021 19:17:36 +0200
From:   Sandor Bodo-Merle <sbodomerle@...il.com>
To:     Ray Jui <rjui@...adcom.com>, Scott Branden <sbranden@...adcom.com>,
        bcm-kernel-feedback-list@...adcom.com,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Sandor Bodo-Merle <sbodomerle@...il.com>
Cc:     Pali Rohár <pali@...nel.org>,
        Ray Jui <ray.jui@...adcom.com>,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: [PATCH] PCI: iproc: restrict multi-MSI to single core CPUs

Commit fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
introduced multi-MSI support with a broken allocation mechanism (it failed to
reserve the proper number of bits from the inner domain).  Natural alignment of
the base vector number was also not guaranteed.

The interrupt affinity scheme used by this driver is incompatible with
multi-MSI as implies moving the doorbell address to that of another MSI group.
This isn't possible for Multi-MSI, as all the MSIs must have the same doorbell
address. As such it is restricted to systems with single CPU core.

Fixes: fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
Reported-by: Pali Rohár <pali@...nel.org>
Signed-off-by: Sandor Bodo-Merle <sbodomerle@...il.com>
---
 drivers/pci/controller/pcie-iproc-msi.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git drivers/pci/controller/pcie-iproc-msi.c drivers/pci/controller/pcie-iproc-msi.c
index eede4e8f3f75..2e42c460b626 100644
--- drivers/pci/controller/pcie-iproc-msi.c
+++ drivers/pci/controller/pcie-iproc-msi.c
@@ -171,7 +171,7 @@ static struct irq_chip iproc_msi_irq_chip = {
 
 static struct msi_domain_info iproc_msi_domain_info = {
 	.flags = MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
-		MSI_FLAG_MULTI_PCI_MSI | MSI_FLAG_PCI_MSIX,
+		MSI_FLAG_PCI_MSIX,
 	.chip = &iproc_msi_irq_chip,
 };
 
@@ -252,18 +252,15 @@ static int iproc_msi_irq_domain_alloc(struct irq_domain *domain,
 
 	mutex_lock(&msi->bitmap_lock);
 
-	/* Allocate 'nr_cpus' number of MSI vectors each time */
-	hwirq = bitmap_find_next_zero_area(msi->bitmap, msi->nr_msi_vecs, 0,
-					   msi->nr_cpus, 0);
-	if (hwirq < msi->nr_msi_vecs) {
-		bitmap_set(msi->bitmap, hwirq, msi->nr_cpus);
-	} else {
-		mutex_unlock(&msi->bitmap_lock);
-		return -ENOSPC;
-	}
+	/* Allocate 'nr_irqs' multiplied by 'nr_cpus' number of MSI vectors each time */
+	hwirq = bitmap_find_free_region(msi->bitmap, msi->nr_msi_vecs,
+					order_base_2(msi->nr_cpus * nr_irqs));
 
 	mutex_unlock(&msi->bitmap_lock);
 
+	if (hwirq < 0)
+		return -ENOSPC;
+
 	for (i = 0; i < nr_irqs; i++) {
 		irq_domain_set_info(domain, virq + i, hwirq + i,
 				    &iproc_msi_bottom_irq_chip,
@@ -284,7 +281,8 @@ static void iproc_msi_irq_domain_free(struct irq_domain *domain,
 	mutex_lock(&msi->bitmap_lock);
 
 	hwirq = hwirq_to_canonical_hwirq(msi, data->hwirq);
-	bitmap_clear(msi->bitmap, hwirq, msi->nr_cpus);
+	bitmap_release_region(msi->bitmap, hwirq,
+			      order_base_2(msi->nr_cpus * nr_irqs));
 
 	mutex_unlock(&msi->bitmap_lock);
 
@@ -539,6 +537,9 @@ int iproc_msi_init(struct iproc_pcie *pcie, struct device_node *node)
 	mutex_init(&msi->bitmap_lock);
 	msi->nr_cpus = num_possible_cpus();
 
+	if (msi->nr_cpus == 1)
+		iproc_msi_domain_info.flags |=  MSI_FLAG_MULTI_PCI_MSI;
+
 	msi->nr_irqs = of_irq_count(node);
 	if (!msi->nr_irqs) {
 		dev_err(pcie->dev, "found no MSI GIC interrupt\n");
-- 
2.31.0